The Helix research group
Research themes
Work in progress and results
Publications
Software and databases
News from Helix
Software and databases
Home page
Site map Mail to Helix
AROM: entity-relationship knowledge modeling
 

AROM ("Allier Relations et Objets pour Modéliser" - "Associate Relationships and Objets for Modeling") is both a knowledge representation formalism and a knowledge base management system that implements this formalism.

AROM thus provides an entity-relationship model, which distinguishes classes and associations between classes. It is somewhat similar to UML (Unified Modeling Language). As in any object-oriented model, a class is described through a set of typed attributes and inherits information from its upper classes in the so-called « is a » hierarchy. An association can link several classes; it is defined by the roles these classes play in the associations and by cardinality constraints; it may have attributes and can be organized in « is a » hierarchies. As a consequence of the existence of associations, an attribute in a class cannot be typed by a class, i.e. attributes always define properties.

The AROM model is very adequate for representing biological data and knowledge. The explicit description of associations allow to design pertinent and easy to understand knowledge bases. Several examples, extracted from real applications, illustrate this point.

As a knowledge base management system, AROM offers several powerful mechanisms. Its graphical interface allows designing and drawing the structure of a knowledge base. Once the structure of classes and associations has been designed, the knowledge base can be populated by creating instances: instances of classes are related through instances of associations. Requests on the contents of the knowledge base can be set. Unknown values of attributes may be computed, either by procedural attachments, or by computing arithmetic expressions. A classification mechanism can determine the lowest classes in the « is a » hierarchy an instance may belong to, thus increasing the information on the related entity.

AROM is written in Java. Its fully documented API makes it easy to integrate in a larger system. AROM is, for example, the very substrate of GenoStar and of GenoExpertBacteria.

To known more on AROM and to freely download the version 2.2.3.

 
In the same section
The GenoStar integrated bioinformatics platform for exploratory genomics
GEB: GenoExpertBacteria
GNA: Genetic Network Analyzer
PepLine: high throughput proteomics
Herbs: checking the consistency of proteome annotations
ISee: In Silico biology e-learning environment
BOX: XML specifications of genomic data
AROM: entity-relationship knowledge modeling
 
    Top of page   Home page  Prepare to print